Current source for 20mA LED?

[Spoerle]

I am rebuilding a machine standing outside.
The power supply will be via 400V (motor 4kW) control electronics over transformer 400V/2x12V 16VA.
One 12V winding is rectified and used to power the MCU.
This voltage also powers the signaling LED.
10mm RGB LED, 20 mA, common +.
Signaling LEDs switch contactor, relays and MCUs via transistors.

The question is whether and how to solve the source for the LED
1. Nothing special, + on 16V DC and limit the current with a resistor to approx. 17 mA. Pros very simple
2. A simple 20mA current source with a resistor and two NPN transistors. Cons complexity of a total of 4 components for every color in LED RGB
3. Current source with LM317LZ and resistor. Cons Price
4. Something else?

Thanks

 

[KlausST]

Hi,

there´s nothing wrong in using juat a resistor. It´s the most used way.

The current neither needs to be very accurate, nor constant. It just needs to be limited below the max specified value for all the expectable cases:
* temperature drift
* supply voltage drift
* and so on

Klaus

 

[Spoerle]

I know it . Under normal circumstances, I wouldn't deal with anything other than resistance.
The devil is in the details.

I have no idea what turning the 4kW motor on/off did to the supply voltage (the contactor for the motor is in the same box)
All this in a temperature range of -10 to +40 C.

I made the MCU power supply from DC +16V, switched stepdown to +5V and then LDO to +3.3V with several LC filters + shielding. +16V DC is used on the switching relay, on the signal LED and the pair of AC12V in series (24V AC) controls the contactor coil.

I really have no idea what limit voltages I should expect on the secondary of the 12AC transformer?

 

[KlausST]

Hi,

i don't understand the whole problem.

Now that you tell you have 5V, then I´d use this 5V ... with a resistor.

If you need more information, please draw a sketch about your situation. And clearly explain what information you need.

Klaus

 

[barry]

You say “the devil is in the details”, but give us no details. Are you for some reason expecting your 5V to jump up some higher voltage? Why?

 

[danadakk]

2 terminal Current Diodes - https://www.diodes.com/assets/Datasheets/AL5809.pdf


Regards, Dana.

 

[crutschow]

The only reason to use constant-current circuit instead of a resistor to drive an LED is if the LED supply voltage significantly varies, and that does not seem to be the case here.

 

[barry]

Overengineering.

Why use a relatively complex $.75 part when a $.02 resistor will work?

 

[danadakk]

1) Brightness variation due to motor start line current surges....

2) Low frequency trigger of seizures due to certain kinds of loads,
for certain sensitive patients, or possibly gerbils......

3) Desire to insure LED is properly and humanely fed correct stable current .....

So many more reasons.....


Regards, Dana.

 

[KlausST]

Hi,

if the 5V stepdown and/or the 3.3V LDO output is not stable enough to drive a LED then it won´t be able to reliably supply any other semiconductor device.
Then the whole system is crap. ... and the LED is the least problem.

Klaus

 

[danadakk]

Huh -

+16V DC is used on the switching relay, on the signal LED and....

Regards, Dana.

 

[dick_freebird]

A right-sized JFET can be a decent current source, very
supply-rejecting although tempco is nobody's friend.

I see a lot of suggestions about "things that could go
wrong" but little to go on regarding probabilities and
consequences (DFMEA style). You might elect to start
simplest and address "problems" when / if they turn out
to be.

 

[KlausST]

Now that you tell you have 5V, then I´d use this 5V ... with a resistor.

It makes no sense to use the 16V. Neither for effort, nor brightness stability, nor power dissipation....

Klaus

 

[Spoerle]

For safety reasons, I do not want the 5V voltage to be on any terminal in the Distribution board.
One signal LED is controlled directly by the auxiliary contacts of the contactor.
If mains voltage were to reach these contacts, it would reliably destroy the board with the MCU. If they were connected to the PE the MCU would stop working, but the power contactor might stay ON.
With 12 and 24V levels, there is no danger of any of this. Connection to PE reliably shuts down the entire machine, Connection to the mains voltage also with the fact that the damage will be precisely defined and the MCU will survive..
if 5V were on the wires in the distribution board it could affect EMI and MCU

 

[KlausST]

Hi,

"Safety reasons"? What is more "unsafe" on 5V than on 16V?

Why should mains voltage reach these contacts? (For safety reasons this must never happen. But still...) And If it reaches, then it will have the same "safety" problem as it reaches 16V. I don´t see any galvanic isolation between 16V and 5V.

Why should it be connected to the PE? And Why should the MCU stop working, and why the power contactor should stay ON?
Why isn´t the same danger with 12V and 24V? Why should a connection to PE shut anything down? How can this be reliably?
"The damage precisely defined"???
"EMI (electromagnetic interferance?) and MCU"???

Wow, that many statements without explanation and any background information. I´ve designed a lot of mains powered devices and can´t find out how your statements can fit to any mains powered device.

*****
In a typical mains powered device:
mains --> transformer --> rectifier --> SMPS 5V --> LDO 3.3V.
.. a single mains failure (one mains side connected to the low voltage side) will have no effect on MCU operation at all. It will just have a safety problem when touching the low voltage side. And it does not matter "where" (24V, 12V, 5V, 3.3V, GND, any other signal) this failure happens.

For an MCU failure you need a second "mains failure" to enable mains current to flow. This most probably will destroy something on the low voltage side. And again it does not matter "where" (24V, 12V, 5V, 3.3V, GND, any other signal) this failure happens: You never can speak of "precisely defined damage".

I don´t know how you come to your conclusions.

But now at post#15 ... it´s time for me to leave. I hope for your project that others have a better understanding of what you are talking about and can better assist you.

Good luck
Klaus

 

[barry]

For safety reasons, I do not want the 5V voltage to be on any terminal in the Distribution board.
One signal LED is controlled directly by the auxiliary contacts of the contactor.
If mains voltage were to reach these contacts, it would reliably destroy the board with the MCU. If they were connected to the PE the MCU would stop working, but the power contactor might stay ON.
With 12 and 24V levels, there is no danger of any of this. Connection to PE reliably shuts down the entire machine, Connection to the mains voltage also with the fact that the damage will be precisely defined and the MCU will survive..
if 5V were on the wires in the distribution board it could affect EMI and MCU

Aren't you also worried about damage from meteors? And what about zombies??? Your concerns seem to make no sense, nor do your assertions that 5V is somehow going to present an EMI issue or endanger the MCU.